netcoredbg exits on macOS after upgrading to netcoredbg 1.2.0-767 #57
Comments
|
Hmm it seems that adding the This, while handling the
Raw messages which trigger this: |
|
OK so the problem there is simple: the values in
That may be just the way I'm running it. I'll investigate |
|
OK well the problem is that Form the help:
But we're using it like this: static span<char> get_exe_path(span<char> buffer)
{
uint32_t real_len;
if (_NSGetExecutablePath(buffer.data(), &real_len) < 0)
return {};
if (real_len >= buffer.size())
return {};
buffer[real_len] = 0;
return buffer.subspan(0, real_len);
}
So:
Also, while this function can return This patch fixes it: diff --git a/src/unix/filesystem_unix.cpp b/src/unix/filesystem_unix.cpp
index ebc0c7e..fbbd78e 100644
--- a/src/unix/filesystem_unix.cpp
+++ b/src/unix/filesystem_unix.cpp
@@ -42,14 +42,11 @@ namespace
#elif defined(__APPLE__)
static span<char> get_exe_path(span<char> buffer)
{
- uint32_t real_len;
+ uint32_t real_len = buffer.size();
if (_NSGetExecutablePath(buffer.data(), &real_len) < 0)
return {};
- if (real_len >= buffer.size())
- return {};
-
- buffer[real_len] = 0;
+ real_len = strlen(buffer.data());
return buffer.subspan(0, real_len);
}
#endif |
|
Hello, @puremourning this looks also wrong fix for me, we need detect size and allocate buffer in this case. As I see, previously code was: this looks like proper logic for me. IMO, we need revert this changes. |
|
Agree 100% |
|
As apple documentation (https://developer.apple.com/library/archive/documentation/System/Conceptual/ManPages_iPhoneOS/man3/dyld.3.html) says: "The bufsize parameter should initially be the size of the buffer". So the real source of the problem, is that |
|
I also see in apple documentation: "With deep directories the total bufsize needed could be more than MAXPATHLEN.". In case of changes proposed by @puremourning we will fail in this case. |
|
Agree with @viewizard - My patch is broken in the case where the returned path > MAXPATHLEN and the code pasted above is not. |
|
@puremourning this issue should be fixed in upstream now. Could you please check it on macOS? |
|
I just updated to Log: |
|
since |
|
Err, you know what, this could be 100% me being a complete fool... (let me just quickly make sure I actually built the app before trying to debug it :/) |
|
Success! I had the wrong framework version on this Mac it turned out. Fixed, and now working! Thanks!
|
|
Nice =) |
|
Fixed in upstream. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi, after upgrading the version of netcoredbg in Vimspector to 1.2.0-767, netcoredbg exits after the initialize exchange. this does not happen with the previous version (netcoredbg 1.2.0-635).
Here is the log including the full message trace:
As you can see, the netcoredbg ("server") exits with status -1 immediately after sending the response to the
launchrequest.With the previous version, this is the (working) log:
After enabling logging in
netcoredbg(with--log) I would that it's throwing an exception and exiting.It seems that the
envkey is now mandatory. this happened before with thecwdkey. Can we make it not mandatory?The text was updated successfully, but these errors were encountered: